The Relationship Between Sea-Swell Bound Wave Height and Wave Shape
Total Page：16
File Type：pdf, Size：1020Kb
Journal of Marine Science and Engineering Article The Relationship between Sea-Swell Bound Wave Height and Wave Shape Floris de Wit * , Marion Tissier and Ad Reniers Faculty of Civil Engineering and Geosciences, Delft University of Technology, 2628CN Delft, The Netherlands; [email protected] (M.T.); [email protected] (A.R.) * Correspondence: [email protected] Received: 9 July 2020; Accepted: 17 August 2020; Published: 21 August 2020 Abstract: The nonlinear wave shape, expressed by skewness and asymmetry, can be calculated from surface elevation or pressure time series using bispectral analysis. Here, it is shown that the same analysis technique can be used to calculate the bound superharmonic wave height. Using measured near-bed pressures from three different field experiments, it is demonstrated that there is a clear relationship between this bound wave height and the nonlinear wave shape, independent of the measurement time and location. This implies that knowledge on the spatially varying bound wave height can be used to improve wave shape-induced sediment transport predictions. Given the frequency-directional sea-swell wave spectrum, the bound wave height can be predicted using second order wave theory. This paper shows that in relatively deep water, where conditions are not too nonlinear, this theory can accurately predict the bispectrally estimated bound superharmonic wave height. However, in relatively shallow water, the mismatch between observed and predicted bound wave height increases significantly due to wave breaking, strong currents, and increased wave nonlinearity. These processes are often included in phase-averaged wind-wave models that predict the evolution of the frequency-directional spectrum over variable bathymetry through source terms in a wave action balance, including the transfer of energy to bound super harmonics. The possibility to calculate and compare with the observed bound super harmonic wave height opens the door to improved model predictions of the bound wave height, nonlinear wave shape and associated sediment transport in large-scale morphodynamic models at low additional computational cost. Keywords: bound wave height; wave shape; bispectrum; field measurements 1. Introduction Coastal management decisions, such as nourishment strategies and sea level rise scenarios, rely more and more on morphodynamic model simulations. Within these simulations, fluxes in sediment transport, caused by hydrodynamic forcing mechanisms, result in changes in the bathymetry. An important contribution to the sediment transport fluxes is the wave shape-induced sediment transport driven by the skewness and asymmetry of the individual waves [1–4]. Although its instantaneous magnitude is often smaller than other contributions, it can have a considerable net effect on the bathymetric evolution as the contribution is typically in the dominant wave direction [5,6]. As such, it is important for beach recovery after storm impact [7,8], onshore bar motion [7,9–11] and the evolution of ebb-tidal shoals (see, e.g., in [12,13]). Current large-scale morphodynamic modeling approaches generally combine a spectral wave transformation model [14–18] and a flow model (see, e.g., in [19–23]) to predict the local wave, flow and sediment transport conditions (see, e.g., in [19,24]). Using a local parameterization based on the wave height, wave period and water depth the wave skewness, asymmetry, and associated sediment transport are obtained (see, e.g., in [25–29]). However, as was shown by Rocha et al. [30], J. Mar. Sci. Eng. 2020, 8, 643; doi:10.3390/jmse8090643 www.mdpi.com/journal/jmse J. Mar. Sci. Eng. 2020, 8, 643 2 of 26 Rocha et al. [31], and De Wit et al. [32], predicting the wave shape using a local approach has its limitations, related to the fact that the prior evolution of wave shape is not taken into account. As a result, the wave shape can be different although the local wave height, period, and water depth are exactly equal, if, for instance, the bed slope is different [33,34], the conditions are rapidly changing [32], or the offshore wave steepness is different [31]. Thus, there is a need for a better way to predict the wave shape that includes the history of the waves before reaching a certain location. The wave skewness and asymmetry can be computed with a bispectral analysis corresponding to the sum of the real and imaginary parts of the bispectrum, respectively (see, e.g., in [35,36]). The bispectrum is a reflection of the coupling between the primary waves and the bound super and sub harmonics [36,37]. This implies that there is a close connection between the nonlinear wave shape and the proportion of bound wave energy. The bound portion of energy in the super harmonics within a directionally spread sea-swell wave field can be predicted with the second order theory of Hasselmann [38] based on a local equilibrium over a horizontal bed. However, in the presence of a variable bathymetry and thus spatially evolving sea-swell wave field this may lead to an erroneous estimate as demonstrated by Herbers and Burton [39]. On the other hand, spectral wind-wave models often include a source term to describe the transfer of wave energy from the primary wind-waves to bound super harmonics through triad sum interactions over variable bathymetry [37,40,41]. The modeled bound fraction of superharmonic wave energy is an integration of the source term in the down-wave direction showing up as an additional spectral peak at twice the primary frequencies (see, e.g., in [37,42]). However, to speed up the computations to enable morphodynamic computations at realistic time scales, the phase information is ignored and even though the spatially evolving fraction of bound energy is implicitly predicted, the accompanying skewness and asymmetry are not known. Examining the three-dimensional (3D) wavenumber-frequency spectrum is a relatively straightforward way to discriminate between the bound and free wave energy as these follow different dispersion relations (see, e.g., in [43]). However, estimating the full 3D wave spectrum requires high-resolution spatial information that is rarely available in the field (see, e.g., in [44] for one of the exceptions). Alternatively, bispectra can be used to characterize the portion of bound energy in a given frequency range. Most efforts to quantify and analyze bound harmonic energy have focussed on the sub-harmonic range (see, e.g., in [45–51]) following the work of Herbers et al. [52] who demonstrated that the bound fraction of subharmonic (i.e., infragravity) energy could be obtained from the difference interactions in the bispectrum. Significantly less attention has been devoted to quantifying the bound energy in the super harmonic range, with the most notable contributions being the work by Herbers and Guza [53,54] and Herbers et al. [55] who examined bound wave energy in intermediate water depths. The aforementioned studies [53–55] showed that triad sum interactions between wave components with large difference angle of propagation can contribute significantly to the bound near-bed pressure variance at these depths. Interestingly, these are typically associated to negative interaction coefficients according to the theory of Hasselmann [38], while the more classical sum interactions between wave components with small difference in angle of propagation have a positive contribution. Thus, for a given sum frequency in the superharmonic range, both positive and negative contributions from primary wave pairs can occur such as the bispectrum is expected to yield a lower limit of the bound super harmonic energy. Several authors [36,56,57] additionally mention, based on the work of McComas and Briscoe [58], that estimating the bound super harmonics from the bispectrum in a broad-banded spectrum is not straightforward. This inhibits a direct comparison with the predictions of the wind-wave spectral models. Notably, Herbers et al. [55] did find a good match between predictions by the theory of Hasselmann [38] and observations in a case of narrow-banded energetic swell conditions (their Figure 10d) in contrast to conditions with crossing sea states (their Figure 10a–c). This raises the question to what extent the bispectral estimate of the bound super harmonic fraction can work for sea-swell conditions. J. Mar. Sci. Eng. 2020, 8, 643 3 of 26 In the following, we therefore construct a method to first estimate the bound portion of the energy in the super harmonics in a realistic directionally spread wave field and secondly to use this as a predictor of the nonlinear wave shape controlling wave skewness and asymmetry. To that end, the velocity and pressure data obtained at nine locations on the Ameland ebb-tidal delta from the CoastalGenesis2/SEAWAD field campaign in September 2017 are examined using bispectral analysis. The bound superharmonic fraction is expressed as an equivalent observed bound wave height that is compared with the predicted bound wave height obtained from the equilibrium theory of Hasselmann [38] to explore its spatial evolution. Next, the correspondence between the bound wave height and nonlinear wave shape is examined to explore the potential of using a wave shape parameterization based on the predicted bound wave height instead of a local parameterization. This is followed up with a discussion on the general applicability of such an approach and the necessary steps in spectral wave modeling to enable these predictions. 2. Background 2.1. The Spectrum The surface elevation is represented as a summation of discrete frequencies as N 1 −2pi fmt 1 ∗ 2pi fmt h(t) = ∑ A( fm)e + A ( fm)e (1) m=−N 2 2 in which A( fm) is the complex amplitude at discrete frequency fm = mD f with D f being the frequency ∗ resolution, A ( fm) indicates the complex conjugate of A( fm), i is the imaginary number, and t is time. The number of discrete spectral estimates is 2N + 1, which are bound by the Nyquist frequencies: ± fN = ± fs/2, in which fs is the discrete sampling frequency of the surface elevation time series.